The $9.3 Billion Procurement List: Embodied Intelligence Enters the Delivery War

A Procurement List That Burst the 'Story Bubble' of Embodied Intelligence

In 2025, the embodied intelligence sector reached a landmark moment — a procurement list totaling 68 billion yuan (approximately $9.3 billion) officially came to light. This is not a number from a product launch PowerPoint, nor a long-term projection from a fundraising roadshow, but a real order written into the books with real money.

This means that, for the first time, embodied intelligence is being asked by buyers to answer the most straightforward questions: What exactly can this money do? How many workers can it replace? How long until we break even?

It took embodied intelligence less than three years to move from the lab to the procurement department. But the real test has only just begun.

Behind the $9.3 Billion: Not a Scale Story, but a Delivery Mandate

Over the past two years, the narrative logic of the embodied intelligence sector could be summed up in a single formula: "Foundation models + robots = infinite possibilities." Capital markets went wild, private market valuations skyrocketed, and players of all stripes unveiled their humanoid robot prototypes — taking a few steps and waving at trade shows was enough to secure another round of funding.

But the emergence of this $9.3 billion procurement list has fundamentally changed the rules of the game.

First, procurement means a delivery commitment. Previously, embodied intelligence companies faced investors and only needed to articulate "what we can do in the future." Now, they face buyers who demand answers to "what can you deliver right now?" Contracts specify delivery timelines, acceptance criteria, and penalty clauses — not technology roadmaps.

Second, a $9.3 billion scale demands mass production capability. No matter how impressive a single humanoid robot demo video may be, it cannot support an order of this magnitude. Buyers don't want a carefully tuned "one-off" from the lab — they want products that can be deployed in volume and run reliably. This places entirely new demands on supply chain management, quality control, and after-sales service systems.

Third, ROI has become the core metric. The deepest signal from this procurement list is that embodied intelligence is being required to "do the math" for the first time. When a robot is deployed on a factory production line or in a logistics warehouse, how many workers can it replace per year? What are the maintenance costs? Is the payback period 18 months or 36 months? These are questions the industry could previously dodge but must now answer head-on.

Who's Paying? Application Scenarios Determine Survival

Based on currently disclosed information, the $9.3 billion procurement covers scenarios spanning manufacturing production lines, warehousing and logistics, specialized operations, and more. This largely aligns with the industry's prior expectations — the first "paying customers" for embodied intelligence are most likely to come from industries with the largest labor shortages and the most standardized operating environments.

Several trends are worth noting:

Industrial scenarios remain the primary battleground. Compared to the complex and variable interaction demands of the service sector, tasks in industrial settings are more structured, and the "unexpected situations" robots need to handle are relatively limited. Material handling, sorting, assembly, and inspection — these high-frequency repetitive tasks are where embodied intelligence can most easily prove its value.

"Humanoid" is no longer a mandatory form factor. Interestingly, not all items on the procurement list point to humanoid robots. Some orders involve "non-humanoid" embodied intelligence devices equipped with environmental perception and autonomous decision-making capabilities. This indicates the market is returning to rationality — form follows function, and customers want solutions, not sci-fi aesthetics.

The role of system integrators is becoming prominent. The era of selling hardware alone may have ended before it even began. Buyers increasingly prefer suppliers that can provide integrated delivery of "robot hardware + deployment solutions + operations and maintenance services." This poses a significant challenge for startups focused solely on building robot bodies.

Industry Watershed: From 'Demo Day' to 'Delivery Day'

The emergence of the $9.3 billion order is accelerating polarization within the industry.

On one side are players that "can deliver." These companies typically share several characteristics: a mature hardware supply chain, algorithms validated in real-world scenarios, and replicable deployment plans. They may not have raised the most funding or published the most technical papers, but they can get robots to customer sites and make them actually "work."

On the other side are players "still telling stories." Their technical approaches have yet to converge, products remain at the prototype stage, and business models rely on fundraising to survive. These companies were already struggling in the funding winter, and as the industry benchmark shifts from "funding raised" to "orders secured," their position will become even more precarious.

One industry insider offered a telling observation: "Investors used to ask how cutting-edge your technology is. Now customers ask whether your robot can work continuously for eight hours without breaking down. The distance between those two questions is the chasm that 90% of embodied intelligence companies have yet to cross."

This chasm is fundamentally a gap in engineering capability. Algorithms that work in the lab may see cliff-like performance drops in real environments when confronted with changing lighting, uneven floors, and irregularly shaped materials. Achieving a 95% success rate in a demo isn't hard, but reaching 99.5% product reliability may require a hundred times more engineering effort.

A Sober Perspective: $9.3 Billion Is a Starting Point, Not the Finish Line

While applauding the industry's progress from hype to substance, we should also remain clear-eyed.

First, orders don't equal revenue. Is the 68 billion yuan a letter of intent or a signed contract amount? Over how many years will it be executed? What are the payment terms? These details will determine the true "gold content" of this order. In recent years, the new energy industry has staged too many dramas of "big orders announced then scaled back," and embodied intelligence should avoid repeating that script.

Second, delivery quality will define the industry's reputation. If the first wave of large-scale embodied intelligence deployments underperforms expectations, the negative impact will extend far beyond any single company. Trust in the entire industry will be damaged, and the difficulty of securing subsequent orders will increase exponentially.

Third, the cost curve still needs to come down. The per-unit cost of current embodied intelligence products remains high, with ROI periods exceeding two years in most scenarios. If costs cannot be reduced by more than 30% within the next 12 to 18 months, scaled deployment will continue to face headwinds.

Outlook: 2025, the 'Coming of Age' for Embodied Intelligence

In retrospect, the $9.3 billion figure itself may not be what matters most. What truly matters is the paradigm shift it represents — the evaluation criteria for the embodied intelligence industry are shifting from "technological potential" to "commercial value," and from "can it be done" to "is it worth buying."

This is the inevitable path every technology sector must walk on the road to maturity. The internet went through it, mobile internet went through it, new energy went through it, foundation models are going through it now, and it's embodied intelligence's turn.

For companies that truly possess delivery capabilities, this is the best of times — the market is finally willing to pay for "products that work" rather than applaud "pretty PowerPoints."

For the industry as a whole, 2025 is likely to be embodied intelligence's "coming of age." Companies that pass this test will truly earn their ticket to a trillion-dollar market; those that don't will become footnotes in this elimination race.

$9.3 billion is not the endgame — it's the opening shot of embodied intelligence's delivery war.